Multi-author manuscripts: practices and experiences of co-authors in the Faculty of Health Sciences, University of the Free State, South Africa

¹Department of Biostatistics, School of Biomedical Sciences, Faculty of Health Sciences, University of the Free State, Bloemfontein, South Africa

^&Corresponding author
Gina Joubert, Department of Biostatistics, School of Biomedical Sciences, Faculty of Health Sciences, University of the Free State, Bloemfontein, South Africa

Introduction    

Multi-author papers are common in health sciences. This may be due to the 'increasing complexity and interdisciplinary nature of research questions' [1]. Jakab et al. [2] reported that from 2000 to 2020, the number of authors on biomedical papers in the NCBI PubMed database increased from a mean of 3.99 authors per publication to a mean of 6.25. In 2000, 17.0% of papers were single-authored, decreasing to 5.7% in 2020 [2]. Some negative connotations are associated with multi-author papers, e.g., 'The perception that co-authors on large authorship publications have not meaningfully contributed underlies widespread institutional bias against multi-authored papers, disincentivizing large collaborations despite their widely recognized value for advancing knowledge' [3]. The International Committee of Medical Journal Editors (ICMJE) recommends that authorship be based on the following four criteria [4]:

Substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work; AND

Drafting the work or revising it critically for important intellectual content; AND

Final approval of the version to be published; AND

Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved

The ICMJE recommendations specifically mention publications based on the work of 'large multi-author groups' and stipulate that all authors of such papers should fulfill all four these criteria [4]. Aim and objectives:

The aim of this study was to review authorship practices and experiences regarding multi-author manuscripts of co-authors in the Faculty of Health Sciences, University of the Free State (UFS), South Africa. The objectives were to provide a profile of qualifying and participating authors in terms of health sciences field and previous multi-author experience

Provide a profile of qualifying and included manuscripts regarding manuscript type, institution of first author, author list format, contributions paragraph and number of authors and countries involved

Summarize multi-authorship practices in terms of authorship discussion, authorship agreements and adherence to ICMJE criteria

Summarize participants´ multi-authorship experiences regarding the included manuscripts and previous multi-author publications

To provide further context the authorship criteria of the journals in which the qualifying manuscripts appeared were summarized.

Methods     

Study design and setting: this cross-sectional study was performed among members of the Faculty of Health Sciences at the University of the Free State, South Africa.

Participants: participants were authors of manuscripts that appeared on the Faculty of Health Sciences research output lists of 2022 and 2023 with 10 or more authors of whom at least one had the Faculty as affiliation but the first author was not affiliated with the Faculty. Manuscripts were excluded if the Faculty author had only an affiliated position and had never been a full-time staff member.

Study size and bias: all qualifying manuscripts were included in the study. The time period of 2022 to 2023 was chosen to limit respondents' recall bias.

Measurements

Faculty authors were contacted by email (via the UFS Outlook address book email or the email provided in the manuscript) in 2024 to introduce the project, and to determine the person's willingness to participate and in which format and when the interview would take place. The information document, consent form and the manuscript(s) that the interview would deal with were attached to this initial email. If no response was received, one reminder email was sent approximately a week later. The author was considered a non-responder if no response was received after the reminder email.

The questionnaire consisted of close-ended questions and some open-ended questions for elaboration and motivation of earlier responses. Close-ended items focused on the participant's fulfillment of the ICMJE authorship criteria, and on the authorship agreement and discussion in the author group. Open-ended items enquired about authorship challenges experienced with the specific manuscript and comments regarding multi-author manuscripts in general.

Structured interviews were conducted by one of the two researchers, either face to face or online with video. Sections of the questionnaire could be answered by a respondent a number of times, depending on their number of qualifying multi-author manuscripts.

Information regarding the number of authors, type of institution of the first author, number of countries of authors and author contributions paragraphs was noted from the published manuscript. Journals' authorship criteria were noted from author or submission guidelines, and journal or ethical policies on journal websites in November 2025. One journal had ceased publication during 2024 and was excluded from this part of the analysis.

Ethical considerations: the protocol was approved by the UFS Health Sciences Research Ethics Committee (HSREC) (UFS-HSD2023/2628/3004-0001), and permission was obtained from applicable UFS authorities. All participants gave written informed consent. All information was treated confidentially and is only reported in summarized form. Interviewees were requested to refrain from identifying any authors by name in their responses. Questionnaires contained only code numbers for manuscripts and respondents.

Pilot study: four Faculty authors (from different Schools or divisions) of qualifying papers of 2023 were approached for the pilot study. Three of these authors responded and participated. No problems were identified with the data collection process, and the pilot study data were included in the main study.

Data analysis: profiles and practices were summarized by frequencies and percentages for categorical variables and medians and ranges for numerical variables. Respondent's; open responses regarding experiences were grouped into categories emanating from the responses, through discussion between the two researchers. Verbatim quotes are indicated by quotation marks.

Results     

Participant profile

Fifty-eight manuscripts published in 47 journals (4 based in South Africa or elsewhere in Africa) qualified for the study. All journals were listed in Web of Science, Scopus or the Directory of Open Access Journals. These manuscripts had 35 Faculty authors of whom 24 (69%) were interviewed. Seven authors (20%) did not respond to the study invitations, one author (3%) was erroneously not contacted and three authors (9%) showed initial interest but did not commit to an interview. Interviews were conducted regarding 32 manuscripts, 55% of the qualifying manuscripts. Respondents were staff members (92%), a staff member who was a postgraduate student (4%). or an affiliate who had been a staff member before (4%). For seven (29%) respondents, this was their first publication with 10 or more authors. The majority of qualifying and participating authors were from the School of Clinical Medicine (Table 1).

Manuscript profile

Table 1 shows the profile of qualifying manuscripts and manuscripts for which interviews were conducted. Compared to qualifying manuscripts, original research manuscripts were overrepresented in the manuscripts for which interviews were conducted, and reviews underrepresented. In addition, manuscripts with first authors from other institutions in South Africa or Africa were overrepresented in the interviews and manuscripts with first authors outside Africa underrepresented compared to all qualifying manuscripts. Contributing to these differences was the fact that one non-responding author had 13 qualifying manuscripts (of which 11 were reviews), all with the first author from an institution outside Africa.

Eleven (19%) qualifying manuscripts included a research group name in the author list, with seven of these listing the research group member names separately and four not stating the names of the research group members (Table 1). There were detailed statements regarding the contributions of the authors named in the author list in 57% of qualifying manuscripts. For all qualifying manuscripts, the median number of authors named in the author lists was 16 (maximum 402), the median number of countries represented in the author list was four (maximum 56), with the median number of total authors 16 (maximum 2 462).

Multi-authorship practices

For three manuscripts, two respondents were interviewed and for four manuscripts, three respondents, thus resulting in 43 unique respondent/manuscript combinations when considering responses relating to a specific manuscript. For 95% of the 43 cases, respondents indicated that they had no difficulty to recall authorship details of the manuscript. Respondents had a median position of 12th author (range 2 to 307).

Respondents reported that authorship was discussed in more than 80% of cases (Table 2) and that a formal authorship agreement was in place for 19% of cases. The most common response (n=20; 47%) to the open-ended question about how authorship was decided on, mentioned only some aspect(s) of criterion 1 of the ICMJE (conception up to data interpretation). For 10 of these cases, the only aspect mentioned was data acquisition. In 22 cases (51%), the respondent indicated that they were not aware how the order of any of the authors or some of the authors was decided on.

Twenty-one respondents (88%) indicated that they were aware of the ICMJE authorship criteria before the interview. Table 3 summarizes respondents' feedback regarding their fulfilment of ICMJE criteria. All respondents indicated that they had made a substantial contribution to at least one aspect of criterion 1. In five cases (12%), the respondent indicated that they had only contributed to the aspect of data acquisition, whereas in 14 cases (33%), respondents indicated that they had contributed substantially to all aspects of criterion 1. In only one case did a respondent indicate that they did not participate in either drafting or critically revising the manuscript (criterion 2). In more than 80% of cases, respondents indicated that they fulfilled all four criteria.

Multi-authorship experiences

For nine of the 43 cases (21%), the respondent mentioned some authorship problems in response to open-ended questions regarding any authorship problems experienced, or any further comment regarding authorship of the specific manuscript. Seven responses related to the extent of the contribution of various authors, and four to differing or problematic approaches or expectations regarding authorship by different institutions. In four cases, the respondent volunteered a positive comment which included 'strong structure, well organised, good trust relationship', 'workshops arranged throughout the process', 'first time to publish in such a high impact journal'. Fifteen (88%) of the 17 respondents who had previous multi-author publications commented on multi-authorship in general (Table 4).

Journals' authorship criteria

Ninety-eight percent (45/46) of journals in which the qualifying manuscripts were published made statements about authorship criteria. Fifty-nine percent (27/46) referred to the ICMJE recommendations regarding authorship, with 21 of these stating specific criteria. Only 12 of these journals stated the four criteria currently recommended by ICMJE correctly. Two additional journals stated the current ICMJE criteria without referring to ICMJE. Fifteen journals (33%) made some reference to group or consortium authorship and five (11%) to multi-author manuscripts.

Discussion     

Previous authorship research in our setting focused on manuscripts with first authors from the Faculty [5]. The previous study included only 6 manuscripts with 10 or more authors, four of which were regarding undergraduate medical student research with all authors from the same institution. To investigate issues related to multi-authorship, the current study therefore included manuscripts with first authors that were not affiliated with our Faculty. Although the manuscript reflects the experiences and practices as reported by co-authors from one institution, the respondents reflect on these aspects regarding interaction with authors from other institutions, which enhances the generalizability of the findings. The decision to use 10 as the number considered as multi-author was somewhat arbitrary, although some journals refer to a maximum number of authors that is considered reasonable. For example, the British Journal of Surgery states that 'most articles are unlikely to involve significant contributions from more than ten authors. If more than ten authors are listed from a single institution, details of their individual contribution will be required' [6].

To provide context to the participants' responses, relevant information regarding the manuscripts and the journals in which they were published was analysed. For just over half of the manuscripts qualifying for the study, detailed statements regarding contributions of authors named in the author list were given. Journals should consider such statements a requirement. As the ICMJE guidelines state, 'it is not the role of journal editors to determine who qualifies or does not qualify for authorship' [4]. However, if journals require author contribution statements, authors would have to consider the various contributions and fulfillment of authorship criteria more seriously. Daly and Texeira da Silva pointed out the importance of clarifying the contributions of middle authors [7].

Nineteen percent of qualifying manuscripts had some form of group authorship. The challenges raised by Hosseini et al. [8] with regard to group authorship, namely 'Who were the group members involved in the project?' and 'What did the group members do?' were also apparent in our study. Andersen et al. [9] have reported the differing ways in which group member names are indicated on manuscripts.

Authorship was discussed among authors in the majority of cases but often this was done only once the project was well under way, namely during data collection or write up. Numbers are small but in the few cases where more than one Faculty author responded regarding the same manuscript, the different authors gave varying responses in this regard, indicating that not all authors were involved in an authorship discussion at the same time. Since authorship can change during the research process and new authors be added [4], a formal authorship agreement (lacking in most cases in this study) as a 'living document' [10] would clarify who qualifies throughout the process and would enable the collective author decisions recommended in this regard by the ICMJE [4].

The majority of respondents indicated that they fulfilled all four of the ICMJE criteria (and were thus closely involved in numerous aspects of the manuscript), whereas their responses regarding the decision about who qualifies as an author reflected less adherence to ICMJE criteria by all authors. Despite their close involvement, approximately half the respondents did not know how the order of author names was decided on. This could be clarified in a formal authorship agreement.

In a 2022 survey of medical physicists in Asia, Europe and North America, participants highlighted 'the quid pro quo effect, division of labor, complimentary expertise, and potential bias against 1-2 author papers' as reasons for multi-author papers [11]. Multi-authorship was also seen as a way to develop professional networks [11]. Borer et al. [3] categorized the meaningful individual contributions in multi-author papers as follows: 1. Intellectual breadth and linkages to new area; 2. Clarity of analytical logic and accuracy; 3. Clarity of logic in the text; 4. Clarity of visual presentation; 5. Assistance with housekeeping' [3]. Respondents in the current study also mentioned some of these positive aspects.

Negative experiences regarding the manuscripts included in the current study centred on the differing contributions of authors, which could be addressed by a formal author agreement. This could also address the concern regarding varying approaches regarding authorship by differing institutions, in particular where not only different institutions but a number of different countries are involved. Haas et al. [12] proposed the following when negotiating authorship: 'communicate often and openly', 'adjust authorship inclusion or position as needed', 'prioritize relationships by acting with respect, empathy, and integrity', and 'address conflict productively' [12].

The majority of journals in which the qualifying manuscripts appeared, referred to ICMJE authorship recommendations in either their submission guidelines or under editorial policies (occasionally with various links that need to be followed to access the information). It is questionable whether authors would always be able to access the information easily or take the trouble to do so. Some journals that referred to the ICMJE authorship recommendations failed to state these criteria, failed to refer to the current four criteria recommended by ICMJE since 2013 [13], or failed to state the criteria correctly.

Limitations

The response rate for authors was close to 70% but only slightly more than 50% of qualifying manuscripts were covered by the interviews. In addition, systematic differences were seen between the manuscripts for which interviews were conducted and all qualifying manuscripts in terms of the type of manuscript, and the location of the first author's institution. These shortcomings and the possibility that authors with less involvement in the manuscript might have chosen not to be interviewed, need to be kept in mind when interpreting the results. For example, one of the authors who initially showed interest but then did not commit to an interview, commented that he did not know why he had been added as an author to the manuscript.

In addition, the varying ways in which institutional affiliations were indicated in manuscripts made it too cumbersome to attempt quantifying the number of institutions represented. In the case of manuscripts with a group name in the author list, affiliations were frequently not stated for the group members. Consequently, we reported only the number of countries represented in the author list.

Recommendations

Authors involved in multi-author manuscripts should familiarize themselves with ICMJE criteria and ensure that authorship is discussed, clarified and documented throughout the research process. Experienced authors should mentor novice authors regarding good authorship practices. Discussions regarding authorship can be held as part of departmental research meetings. Official standpoints regarding multi-authorship by national research authorities could strengthen authors' position in discussions with authors from other countries.

Journal editors should ensure that their authorship guidelines are easily accessible and clear, that they adhere to international guidelines and that they are applied. They should also require that author contributions are indicated and provide guidance regarding the content of these. At our institution, good practices and experiences as well as challenges regarding multi-author publications should be highlighted in the Faculty authorship guidelines.

Conclusion     

Adherence to ICMJE criteria by participating Faculty of Health Sciences co-authors of multi-author manuscripts was high. Formal authorship agreements were rare and in approximately half of the cases, the respondents did not know how the order of author names were decided on. Of the respondents with previous experience of multi-author manuscripts, more mentioned negative rather than positive aspects of such manuscripts. Although the majority of journals in which qualifying manuscripts were published referred to ICMJE criteria, less than half of those stated the current four criteria correctly.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Gina Joubert was responsible for conceptualization, methodology, investigation, validation, software, formal analysis, data curation, project administration, visualization and writing - original draft. Cornel van Rooyen was responsible for methodology, investigation, validation and writing - review and editing. All the authors read and approved the final version of the manuscript.

Acknowledgments     

We thank the participants for their frank responses, and Dr. Daleen Struwig, medical writer/editor, Faculty of Health Sciences, University of the Free State, for language editing and assistance with technical preparation of the article.

Tables     

Table 1: profile of qualifying and participating authors and manuscripts, Faculty of Health Sciences, University of the Free State, 2022-2023

Table 2: respondents' (n=43) feedback regarding authorship discussions and authorship agreement, Faculty of Health Sciences, University of the Free State, 2022-2023

Table 3: respondents' (n=43) feedback regarding their fulfilment of ICMJE authorship criteria, Faculty of Health Sciences, University of the Free State, 2022-2023

Table 4: positive and negative comments regarding multi-author publications of respondents with previous multi-author publications (n=17), Faculty of Health Sciences, University of the Free State, 2024

References